Does Greater Long-term Intraocular Pressure (IOP) Variability Increase Probability Of Developing Primary Open Angle Glaucoma (POAG) in the Ocular Hypertension Treatement Study (OHTS)?

Abstract

Purpose: : To determine whether greater long-term IOP variability among observation participants increases the probability of developing POAG.

Methods: : Observation participants needed to have at least 3 study untreated IOPs to be included in this analysis. Univariate and multivariate Cox proportional hazards models were used to assess measures of IOP variability (standard deviation [SD], maximum [MAX], and range) on the probability of developing POAG. Data were censored after initiation of topical hypotensive medication. IOP variables were categorized into quartiles due to non-linearity. Multivariate (MV) models included baseline age, central corneal thickness (CCT), vertical cup to disc ratio (VCD), pattern standard deviation (PSD), mean follow-up (f/up) IOP (instead of baseline IOP) and a measure of IOP variability. C-statistic and calibration chi-square were used to asses predictive accuracy of models.

Results: : 696 of 819 participants randomized to observation had ≥ 3 untreated IOP measurements for inclusion in this analysis. The sample included 90 incident cases of POAG. Mean f/up IOP was 24.1 mmHg + 3.1 and mean duration of f/up was 6.4 years + 1.4. In univariate models, IOP SD (p=0.02) and MAX (p<0.0001) were statistically significant, range was not statistically significant (p=0.55). In MV models which included baseline age, CCT, VCD, PSD and mean f/up IOP, none of the measures of IOP variability was statistically significant (SD p=0.25, MAX p=0.98, range p=0.25). The c-statistic and calibration X² for the MV prediction model without IOP variability were 0.82 and 6.41, respectively. The addition of IOP SD or MAX or range to this model did not improve the predictive accuracy (c-statistic of 0.82, 0.82 and 0.82, respectively and calibration X² of 6.91, 6.14 and 2.24, respectively).

Conclusions: : Measures of IOP variability (SD, MAX, range) did not improve the predictive accuracy of the MV prediction model (baseline age, CCT, VCD, PSD and mean f/up IOP) in this sample.

Clinical Trial: : http://www.clinicaltrials.gov nct00000125

Keywords: clinical (human) or epidemiologic studies: risk factor assessment